We are investigating delays in processing queued requests for ytr-ai/qwen/image/t2i model API